Book a Demo!
CoCalc Logo Icon
StoreFeaturesDocsShareSupportNewsAboutPoliciesSign UpSign In
Azure
GitHub Repository: Azure/Azure-Sentinel-Notebooks
 Path: blob/master/tutorials-and-examples/feature-tutorials/Base64Unpack.ipynb
3253 views
Kernel: Python 3

Title: msticpy - Base64 Decoder

Description:

This module allows you to extract base64 encoded content from a string or columns of a Pandas DataFrame. The library returns the following information:

  • decoded string (if decodable to utf-8 or utf-16)

  • hashes of the decoded segment (MD5, SHA1, SHA256)

  • string of printable byte values (e.g. for submission to a disassembler)

  • the detected decoded file type (limited)

If the results of the decoding contain further encoded strings these will be decoded recursively. If the encoded string appears to be a zip, gzip or tar archive, the contents will be decompressed after decoding. In the case of zip and tar, the contents of the archive will also be checked for base64 encoded content and decoded/decompressed if possible.

You must have msticpy installed to run this notebook:

%pip install --upgrade msticpy

# Imports
import sys
MIN_REQ_PYTHON = (3,6)
if sys.version_info < MIN_REQ_PYTHON:
    print('Check the Kernel->Change Kernel menu and ensure that Python 3.6')
    print('or later is selected as the active kernel.')
    sys.exit("Python %s.%s or later is required.\n" % MIN_REQ_PYTHON)


from IPython.display import display
import pandas as pd

# Import Base64 module
from msticpy.nbtools import *
from msticpy.sectools import *
mpmagics 
# Load test data
process_tree = pd.read_csv('data/process_tree.csv',
                           parse_dates=["TimeGenerated"],
                           infer_datetime_format=True)
process_tree[['CommandLine']].head()

Contents

Decoding Base64 String

Base64 decode an input string.

    Base64 decode an input string.

    Parameters
    ----------
    input_string : str, optional
        single string to decode (the default is None)
    trace : bool, optional
        Show additional status (the default is None)

    Returns
    -------
    Tuple[str, Optional[List[BinaryRecord]]]
        Decoded string and additional metadata

    Notes
    -----
    Items that decode to utf-8 or utf-16 strings will be returned as decoded
    strings replaced in the original string. If the encoded string is a
    known binary type it will identify the file type and return the hashes
    of the file. If any binary types are known archives (zip, tar, gzip) it
    will unpack the contents of the archive.
    For any binary it will return the decoded file as a byte array, and as a
    printable list of byte values. If the input is a string the function
    returns:

    - decoded string: this is the input string with any decoded sections
      replaced by the results of the decoding

# get a commandline from our data set
cmdline = process_tree['CommandLine'].loc[39]
cmdline
'.\\powershell -enc JAB0ACAAPQAgACcAZABpAHIAJwA7AA0ACgAmACAAKAAnAEkAbgB2AG8AawBlACcAKwAnAC0ARQB4AHAAcgBlAHMAcwBpAG8AbgAnACkAIAAkAHQA'
# Decode the string
base64_dec_str = base64.unpack(input_string=cmdline)

# Print decoded string
print(base64_dec_str)
(".\\powershell -enc <decoded type='string' name='[None]' index='1' depth='1'>$\x00t\x00 \x00=\x00 \x00'\x00d\x00i\x00r\x00'\x00;\x00\r\x00\n\x00&\x00 \x00(\x00'\x00I\x00n\x00v\x00o\x00k\x00e\x00'\x00+\x00'\x00-\x00E\x00x\x00p\x00r\x00e\x00s\x00s\x00i\x00o\x00n\x00'\x00)\x00 \x00$\x00t\x00</decoded>", reference original_string file_name \ 0 (, 1., 1) JAB0ACAAPQAgACcAZABpAHIAJwA7AA0ACgAmACAAKAAnAE... unknown file_type input_bytes \ 0 None b"$\x00t\x00 \x00=\x00 \x00'\x00d\x00i\x00r\x0... decoded_string encoding_type \ 0 $t = 'dir';\r\n& ('Invok... utf-8 file_hashes \ 0 {'md5': '6cd1486db221e532cc2011c9beeb4ffc', 's... md5 sha1 \ 0 6cd1486db221e532cc2011c9beeb4ffc 6e485467d7e06502046b7c84a8ef067cfe1512ad sha256 \ 0 d3291dab1ae552b91e6b50d7460ceaa39f6f92b2cda433... printable_bytes 0 24 00 74 00 20 00 3d 00 20 00 27 00 64 00 69 0... )

Contents

Using a DataFrame as Input

You can replace to base64.unpack_df() to pass a DataFrame as an argument. Use the column parameter to specify which column to process.

In the case of DataFrame input, the output DataFrame contains these additional columns:

  • src_index - the index of the row in the input dataframe from which the data came.

  • full_decoded_string - the full decoded string with any decoded replacements. This is only really useful for top-level items, since nested items will only show the 'full' string representing the child fragment.

Base64 decode strings taken from a pandas dataframe.

Parameters
----------
data : pd.DataFrame
    dataframe containing column to decode
column : str
    Name of dataframe text column
trace : bool, optional
    Show additional status (the default is None)

Returns
-------
pd.DataFrame
    Decoded string and additional metadata in dataframe

Notes

Items that decode to utf-8 or utf-16 strings will be returned as decoded strings replaced in the original string. If the encoded string is a known binary type it will identify the file type and return the hashes of the file. If any binary types are known archives (zip, tar, gzip) it will unpack the contents of the archive. For any binary it will return the decoded file as a byte array, and as a printable list of byte values.

# specify the data and column parameters
dec_df = base64.unpack_df(data=process_tree, column='CommandLine')
dec_df

Contents

Interpreting the DataFrame output.

For simple strings the Base64 decoded output is straightforward. However for nested encodings this can get a little complex and difficult to represent in a tabular format.

Columns

  • reference - The index of the row item in dotted notation in depth.seq pairs (e.g. 1.2.2.3 would be the 3 item at depth 3 that is a child of the 2nd item found at depth 1). This may not always be an accurate notation - it is mainly use to allow you to associate an individual row with the reference value contained in the full_decoded_string column of the topmost item).

  • original_string - the original string before decoding.

  • file_name - filename, if any (only if this is an item in zip or tar file).

  • file_type - a guess at the file type (this is currently elementary and only includes a few file types).

  • input_bytes - the decoded bytes as a Python bytes string.

  • decoded_string - the decoded string if it can be decoded as a UTF-8 or UTF-16 string. Note: binary sequences may often successfully decode as UTF-16 strings but, in these cases, the decodings are meaningless.

  • encoding_type - encoding type (UTF-8 or UTF-16) if a decoding was possible, otherwise 'binary'.

  • file_hashes - collection of file hashes for any decoded item.

  • md5 - md5 hash as a separate column.

  • sha1 - sha1 hash as a separate column.

  • sha256 - sha256 hash as a separate column.

  • printable_bytes - printable version of input_bytes as a string of \xNN values

Contents

SourceIndex column allows you to merge the results with the input DataFrame

Where an input row has multiple decoded elements (e.g. a nested encoding or a zip or other archive file), the output of this merge will result in duplicate rows from the input (one per element match). The DataFrame index from the source is preserved in the src_index column.

Note: you may need to force the type of the src_index column to be the same type as the original DataFrame in order to merge. In the example below case we are matching with the default numeric index so we force the type to be numeric. In cases where you are using an index of a different dtype you will need to convert the src_index (dtype=object) to match the type of your index column.

# Set the type of the SourceIndex column. 
dec_df['SourceIndex'] = pd.to_numeric(dec_df['src_index'])
merged_df = (process_tree
             .merge(right=dec_df, how='left', left_index=True, right_on='SourceIndex')
             .drop(columns=['Unnamed: 0'])
             .set_index('SourceIndex'))

# Show the result of the merge (only those rows that have a value in original_string)
merged_df.dropna(subset=['original_string'])

# Note the output of unpack_items() may have multiple rows (for nested encodings) 
# In this case merged DF will have duplicate rows from the source.

Contents

Decoding Nested Base64/Archives

The module will try to follow nested encodings. It uses the following algorithm:

  1. Search for a pattern in the input that looks like a Base64 encoded string

  2. If not a known undecodable_string, try to decode the matched pattern.

    • If the base 64 string matches a known archive type (zip, tar, gzip) also decompress or unpack

      • For multi-item archives (zip, tar) process each contained item recursively (i.e. go to item 1. with child item as input)

    • For anything that decodes to a UTF-8 or UTF-16 string replace the input pattern with the decoded string

    • Recurse over resultant output (i.e. submit decoded/replaced string to 1.)

  3. If decoding fails, add to list of undecodable_strings (prevents infinite looping over something that looks like a base64 string but isn't)

encoded_cmd = '''
powershell.exe  -nop -w hidden -encodedcommand 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'''

import re
dec_string, dec_df = base64.unpack(input_string=encoded_cmd)
print(dec_string.replace('<decoded', '\n<decoded'))
powershell.exe -nop -w hidden -encodedcommand <decoded value='multiple binary' type='multiple' index='1' depth='1'> <decoded value='binary' name='[zip] Filename: PeDll.dll' type='zip' index='1.1' depth='2'>3e 7b f4 bf 50 41 33 30 01 23 9d 3f 8d 4c d4 01 b0 5e 08 d0 3f c4 0c 01 a0 71 00 50 08 21 9c a6 12 1a 66 81 4b 3f a9 a6 d3 9e 53 60 80 22 01 03 00 00 80 00 00 00 00 00 00 00 00 18 c0 83 f6 fc 01 60 2d aa aa aa aa aa aa aa aa aa aa aa 0a aa aa 1a 1a 80 a1 2d aa aa aa aa aa aa aa aa aa 2a a2 11 fa c8 00 e8 7f 01 60 fd 07 c0 ff 05 80 ff 07 c0 ff 05 40 ff 01 46 00 b3 03 40 28 87 91 69 76 00 c8 20 a3 03 00 20 62 13</decoded> <decoded type='string' name='[zip] Filename: b64inzip.foo' index='1.2' depth='2'>Base64 encoded string in zip file</decoded> <decoded type='string' name='[zip] Filename: PlainTextInZip.dll' index='1.3' depth='2'>Unencoded text file in zip</decoded></decoded>

IPython magic

You can use the line magic %b64 or cell magic %%b64 to decode text directly in a cell

The b64 magic supports the following options:

--out OUT, -o OUT  The variable to return the results in the variable `OUT`
                   Note: the output is a tuple of decoded string and pandas DataFrame
--pretty, -p       Print formatted version of output (if you `print` the output)
--clean, -c        Print decoded string with no formatting

%%b64 --pretty --out dec_xml
powershell.exe  -nop -w hidden -encodedcommand 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'<?xml version="1.0" encoding="utf-8"?>\n<decoded_string>\n powershell.exe -nop -w hidden -encodedcommand\n <decoded depth="1" index="1" type="multiple" value="multiple binary">\n <decoded depth="2" index="1.1" name="[zip] Filename: PeDll.dll" type="zip" value="binary">\n 3e 7b f4 bf 50 41 33 30 01 23 9d 3f 8d 4c d4 01 b0 5e 08 d0 3f c4 0c 01 a0 71 00 50 08 21 9c a6 12 1a 66 81 4b 3f a9 a6 d3 9e 53 60 80 22 01 03 00 00 80 00 00 00 00 00 00 00 00 18 c0 83 f6 fc 01 60 2d aa aa aa aa aa aa aa aa aa aa aa 0a aa aa 1a 1a 80 a1 2d aa aa aa aa aa aa aa aa aa 2a a2 11 fa c8 00 e8 7f 01 60 fd 07 c0 ff 05 80 ff 07 c0 ff 05 40 ff 01 46 00 b3 03 40 28 87 91 69 76 00 c8 20 a3 03 00 20 62 13\n </decoded>\n <decoded depth="2" index="1.2" name="[zip] Filename: b64inzip.foo" type="string">\n Base64 encoded string in zip file\n </decoded>\n <decoded depth="2" index="1.3" name="[zip] Filename: PlainTextInZip.dll" type="string">\n Unencoded text file in zip\n </decoded>\n </decoded>\n</decoded_string>'
print(dec_xml[0])
<?xml version="1.0" encoding="utf-8"?> <decoded_string> powershell.exe -nop -w hidden -encodedcommand <decoded depth="1" index="1" type="multiple" value="multiple binary"> <decoded depth="2" index="1.1" name="[zip] Filename: PeDll.dll" type="zip" value="binary"> 3e 7b f4 bf 50 41 33 30 01 23 9d 3f 8d 4c d4 01 b0 5e 08 d0 3f c4 0c 01 a0 71 00 50 08 21 9c a6 12 1a 66 81 4b 3f a9 a6 d3 9e 53 60 80 22 01 03 00 00 80 00 00 00 00 00 00 00 00 18 c0 83 f6 fc 01 60 2d aa aa aa aa aa aa aa aa aa aa aa 0a aa aa 1a 1a 80 a1 2d aa aa aa aa aa aa aa aa aa 2a a2 11 fa c8 00 e8 7f 01 60 fd 07 c0 ff 05 80 ff 07 c0 ff 05 40 ff 01 46 00 b3 03 40 28 87 91 69 76 00 c8 20 a3 03 00 20 62 13 </decoded> <decoded depth="2" index="1.2" name="[zip] Filename: b64inzip.foo" type="string"> Base64 encoded string in zip file </decoded> <decoded depth="2" index="1.3" name="[zip] Filename: PlainTextInZip.dll" type="string"> Unencoded text file in zip </decoded> </decoded> </decoded_string>

Pandas Extension

The decoding functionality is also available in a pandas extension mp_b64. This supports a single method extract().

This supports the same syntax as unpack_df (described earlier).

process_tree.mp_b64.extract(column='CommandLine')

Contents

To-Do Items

  • Use more comprehensive list of binary magic numbers and match on byte values after decoding to get better file typing

  • Output nested decodings in a more readable output